A method is presented for using multiple protein structures (MPS) in computational drug design. The initial goal of this proposal is to provide a solid technique for incorporating protein flexibility into structure-based drug discovery (SBDD). The modeling community needs methods that address this long-standing problem, and the task provides a measure of the strengths and weaknesses of using MPS before tackling the greater challenge of structural genomics. NIGMS has recently funded seven research centers as part of its Protein Structure Initiative. Drug discovery will be revolutionized with the explosion of information to come, but only if the tools exist to combine many related protein structures in a way that is useful for SBDD. The long-term goal of this work is to improve the field of computer-aided drug design by developing methods that more accurately model target proteins and exploit the vast amount of information available from proteomics. Better methods for drug design will speed the discovery of lead compounds and new pharmaceutical therapies. The specific aims for this project focus on (1) optimizing the protocol for using MPS from computer simulations and experimental sources, (2) extrapolating the use of MPS to exploit a subfamily of homologous proteins in the development of broad-spectrum therapeutics, and (3) applying the method to systems of biomedical importance. MPS will be used to create receptor-based pharmacophore models for several enzymatic systems. The models will be judged by searching a database of active and inactive inhibitory compounds from the literature. Successful models will present few false positives and will identify the most active compounds in the database. Collaborations with enzymologists are proposed to aid the development of novel inhibitors in the later stages of the project.